Re: [PATCH v5] lockdep: Allow tuning tracing capacity constants.

[Tetsuo Handa]

Peter, are you there?

If you keep silence, we will assume that applying this patch is the way to go.

On 2021/03/20 16:34, Dmitry Vyukov wrote:
> On Mon, Feb 8, 2021 at 11:29 AM Tetsuo Handa
> <penguin-kernel@xxxxxxxxxxxxxxxxxxx> wrote:
>>
>> Since syzkaller continues various test cases until the kernel crashes,
>> syzkaller tends to examine more locking dependencies than normal systems.
>> As a result, syzbot is reporting that the fuzz testing was terminated
>> due to hitting upper limits lockdep can track [1] [2] [3]. Since analysis
>> via /proc/lockdep* did not show any obvious culprit [4] [5], we have no
>> choice but allow tuning tracing capacity constants.
>>
>> [1] https://syzkaller.appspot.com/bug?id=3d97ba93fb3566000c1c59691ea427370d33ea1b
>> [2] https://syzkaller.appspot.com/bug?id=381cb436fe60dc03d7fd2a092b46d7f09542a72a
>> [3] https://syzkaller.appspot.com/bug?id=a588183ac34c1437fc0785e8f220e88282e5a29f
>> [4] https://lkml.kernel.org/r/4b8f7a57-fa20-47bd-48a0-ae35d860f233@xxxxxxxxxxxxxxxxxxx
>> [5] https://lkml.kernel.org/r/1c351187-253b-2d49-acaf-4563c63ae7d2@xxxxxxxxxxxxxxxxxxx
>>
>> Reported-by: syzbot <syzbot+cd0ec5211ac07c18c049@xxxxxxxxxxxxxxxxxxxxxxxxx>
>> Reported-by: syzbot <syzbot+91fd909b6e62ebe06131@xxxxxxxxxxxxxxxxxxxxxxxxx>
>> Reported-by: syzbot <syzbot+62ebe501c1ce9a91f68c@xxxxxxxxxxxxxxxxxxxxxxxxx>
>> References: https://lkml.kernel.org/r/1595640639-9310-1-git-send-email-penguin-kernel@xxxxxxxxxxxxxxxxxxx
>> Signed-off-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
>> Acked-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
> 
> Peter, ping.
> Please at least provide some feedback.
> This is still the top crasher on syzbot and harms testing of the whole
> kernel. I am periodically thinking of disabling LOCKDEP on syzbot as
> harming more than helping, but so far talking myself out of it because
> it will likely be broken more when we try to re-enable it and I still
> hope for a timely resolution of this issue.
>